Transmission Electron Microscopy as the Visualization Technique for Analysis of Circadian Synaptic Plasticity in the Mouse Barrel Cortex

Summary

The presented protocol describes the use of transmission electron microscopy (TEM) to quantify circadian changes in the mouse barrel cortex, mainly focusing on synapse number and dendritic spine morphology.

Abstract

Examining circadian synaptic plasticity requires housing mice under different lighting conditions (light/dark cycle, LD 12:12, and constant darkness, DD), providing access to running wheels, and sacrificing them at four defined time points within 24 h-at the beginning and middle of the day/subjective day and at the beginning and middle of the night/subjective night. Brains are then properly fixed for transmission electron microscopy (TEM). The barrel cortex, with its precise somatotopic organization, provides an ideal model for such analysis. To obtain the required brain area, the brains are tangentially cut with a vibratome, and then, sections containing the barrel cortex are selected and embedded in Polybed resin. From the prepared blocks containing the selected barrels, consecutive ultrathin sections are cut. Synaptic density, excitatory and inhibitory, is analysed from electron micrographs using the stereological dissector method. Additionally, TEM images are used for 3D reconstructions of dendritic spines. Changes in the shape of dendritic spines indicate remodeling of neurons during the day. The number of excitatory synapses peaks during sleep (day) in mice, while inhibitory synapses peak during their activity phase (in the middle of the night).

Introduction

Circadian rhythms are generated by circadian clocks in almost all processes in an organism. In animals and humans, they have been detected at molecular, cellular, and whole-organism levels, as well as in their behavior. The circadian system of an organism consists of the main circadian clock (pacemaker) and peripheral clocks. All circadian clocks generate circadian oscillations through the cyclic expression of clock genes, which are controlled by their proteins. The molecular mechanism of the clock generates circadian rhythms with a period of ~1 day (longer or shorter than 24 h), but under day/night conditions, the period of endogenous rhythms is synchronized to 24 h. Many circadian rhythms have been detected in the physiology of the nervous system of invertebrates and vertebrates; however, some studies have shown that they are also present in synaptic and neuronal plasticity, that is, changes in the number and structure of synapses and dendrites^1,2,3,4.

The mouse somatosensory cortex provides an excellent site for studying synaptic plasticity throughout the day due to its well-defined organization and direct link to animal locomotor activity^5,6,7,8,9. In layer IV of the somatosensory cortex, there are distinct neural structures (barrels) that are highly noticeable even in an unstained brain^10,11. Their arrangement reflects the morphology and organization of the whiskers on the animal's snout, allowing the precise mapping of sensory stimuli. The large barrels are organized in five regular rows, each containing between four and seven barrels. The rows of barrels that correspond to large whiskers are marked with capital letters from A to E and represent the whiskers arranged from the eye area downward along the snout. Each barrel consists of two main areas: the hollow, the central part filled with neuropil, and the wall called the side, which is dominated by the bodies of stellate cells. The barrels are separated from each other by areas of lower cell density, called septa. Each barrel receives impulses from a specific whisker located on the contralateral side of the animal's snout^10,12.

The division into specific barrels, arranged in regular rows, allows for quick and easy identification of the selected area^10,13. The selection of barrels from row B for analysis of circadian changes is based on studies of activity-dependent plasticity^14,15,16,17. Row B is characterized by well-defined somatotopic boundaries and contains only four distinct large barrels, simplifying its location in samples. Due to its central position, row B neuronal activity is correlated with natural mouse behaviors such as environmental exploration and locomotor activity. For counting synapses on dendritic spines and shafts, it is important to select the central parts of the barrels, where cell bodies are sparse^12,15,18.

Immunohistochemistry combined with confocal microscopy imaging is a technique that has played a crucial role in clarifying the key mechanisms of synaptic plasticity in mice. While using confocal microscopy provides high-resolution images with reduced background fluorescence and improved cellular structure clarity, it suffers from photobleaching and requires complex sample preparation, which may change the state of biological tissues. Its reliance on fluorescent markers can limit the types of proteins that can be studied simultaneously due to their overlap. Confocal immunofluorescence microscopy provides a means to label specific proteins, their expression levels, and localization associated with synaptic activity by using antibodies to detect target antigens in fixed tissue sections. Although this method offers specificity and flexibility to study various proteins involved in neurotransmission, it is limited by its inability to provide real-time data on dynamic processes due to the fixation step involved. The inconsistency in the antibody binding could also result in unreliable results^19,20,21.

The use of electron microscopy instead of immunofluorescence methods with light microscopy allows for significantly higher resolution and greater precision in both localization selection and the distinction of neural structures. Serial transmission electron microscopy (TEM) sections help to visualize the synaptic structure at the ultrastructural level and have the advantage of being used for studying synapse formation or elimination. TEM enables researchers to analyze the physical properties of synapses and changes in the density and structure that are called the plasticity of synapses. Although sample preparation and imaging are labor-intensive, TEM provides a more detailed understanding of changes that happen during learning, memory formation, and processing of sensory information, and during the day, year, and animal age.

Our method aims to investigate the circadian dynamics of synaptic density and morphology by using a minimum four-time-point-based approach combined with the stereological dissector method based on serial TEM images. The stereological dissector method allows for reliable estimation of synapse numbers even when using only a few ultrathin serial sections^{22,23,24,25,26}. The well-defined somatotopic organization of the barrel cortex ensures precise anatomical selection of study locations, while serial TEM allows differentiation of synapse types (excitatory and inhibitory) and their locations (on dendritic spines and shafts). This method provides a valuable tool for researchers investigating diurnal and circadian neuroplasticity, enabling the exploration of how environmental and internal factors influence synaptic dynamics. The individual components of our method enable its application for analysis of neuroplasticity changes in a much broader range of studies, not necessarily limited to circadian-related changes.

Protocol

All experimental procedures involving animals were approved by the appropriate institutional ethics committee and conducted in accordance with Directive 2010/63/EU of the European Parliament and of the Council on the protection of animals used for scientific purposes, as well as with national regulations. All efforts were made to minimize animal suffering and to reduce the number of animals used.

1. Preparation of brain tissues

1. Sacrifice mice at four different time points every 6 h over a 24 h cycle, in both light/dark (LD 12:12) and constant darkness (DD) conditions (Figure 1). In LD 12:12 conditions, refer to Zeitgeber Time (ZT), where ZT0 corresponds to the beginning of the light phase and ZT12 to the beginning of the dark phase. In constant darkness, refer to Circadian Time (CT), where CT0 represents the beginning of the subjective day, and CT12 denotes the beginning of the subjective night.
   NOTE: The time points are ZT0/CT0 (start of the day, subjective day), ZT6/CT6 (middle of the day, subjective day), ZT12/CT12 (beginning of the night, subjective night), and ZT18/CT18 (middle of the night, subjective night).
2. Perfuse and fix mouse brains with 2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer (pH 7.4).
3. Remove each brain from the skull and leave it in a fixative solution (2.5% glutaraldehyde and 2% paraformaldehyde in 0.1 M phosphate buffer, pH 7.4) for 24 h at 4 °C.
4. To prepare the brain for sectioning, cut off the cerebellum and divide the brain into two hemispheres (Figure 2A,B).
5. Choose one hemisphere (right or left) and orient it so that the sections are cut tangentially to the barrel cortex. Mount the selected hemisphere on the vibratome chuck and glue it tightly (Figure 2C).
6. Cut sections at 60 µm thickness (vibratome section speed: 7.10 and the frequency: 5.00) (Figure 2D) and then transfer sections to the mix of 0.1 M phosphate buffer and fixative solution (1:5 ratio).
7. Examine sections under a light microscope with 4x magnification and collect only those with the visible barrel field cortex for the next step (Figure 2E). Use the concave slide to ensure that the section remains constantly in liquid.

2. Brain fixation and sectioning

1. Use a paintbrush and gently transfer tissues into a small glass Petri dish. Next, rinse the brain sections in the 0.1 M cacodylate buffer (BC, pH 7.4) for 3 x 5 min.
2. Fix sections in 1% Osmium tetroxide (OsO4) in 0.1 M BC with 1.5% potassium ferricyanide and leave at +4 °C for 1 h.
   NOTE: Prepare the OsO4 solution immediately before use since OsO₄ is unstable and its activity decreases over time.
3. Post-fix sections in 1% OsO4 in 0.1 M BC solution for 1 h at room temperature.
   NOTE: Perform all these steps while wearing protective gloves and under the fume hood with well-functioning ventilation as OsO4 is very toxic.
   Waste should be kept in a separate bottle with a tight cap. Do not allow the product to enter drains.
4. After fixation, rinse sections in distilled water for 2 x 5 min.
5. In the next step, incubate brain sections for 40 min in the solution of 70% ethanol containing 1% uranyl acetate. After mixing, use a syringe with a filter (25 mm) and slowly pour 2 mL of the solution into each Petri dish with the tissues.
6. Dehydrate tissues in ascending concentrations of alcohol: 70%, 80%, and 90% ethanol for 5 min each and 100% ethanol, 3 x 5 min each.
   NOTE: It is possible to leave tissues in 70% ethanol overnight at +4 °C.
7. Wash tissues in propylene oxide 2 x 10 min.
   NOTE: Propylene oxide is a highly corrosive chemical; use it with caution while wearing protective gloves and working under the fume hood. Using a glass container (glass Petri dish) is crucial for this step.
   Propylene oxide is an extremely flammable liquid and vapor. Waste must be kept in a separate tightly closed glass container. Do not let the product enter drains; there is a risk of explosion.
8. Replace propylene oxide with a mixture of Polybed resin and propylene oxide (1:1 ratio). Pour at least 2 mL of the solution into each Petri dish with the sections, cover with a lid, and incubate for 40 min.
9. Finally, embed brain sections in the mixture of Polybed resin and propylene oxide (3:1 ratio). Use at least 2 mL of the mixture per Petri dish, cover it and leave for 1.5 h.
   NOTE: Ensure the tissues are fully immersed and make sure they are not floating above the liquid.
10. Cut Aclar film into pieces matching the size of a glass slide. Using a plastic pipette, apply a small volume of resin onto the Aclar film and transfer the brain sections from the Petri dish to the resin with a paintbrush. Embed each section between two Aclar films and incubate for 48 h in a 65 °C oven to polymerize.

3. Barrel cortex imaging and sectioning

1. Photograph brain sections at 2x objective magnification with a light microscope (Figure 2F).
2. Collect all images and import them into graphics software.
   1. To open images, click File |Open, and select the images using Shift |Open. The images will open in separate windows.
   2. Start stacking images from the cortex (the smallest piece) and continue until all sections are correctly aligned.
   3. Right-click on Layer 1 |Duplicate Layer on each image. In the dialog box, enter the image number as the name, choose one image as the destination file, and click OK. Repeat these steps for all the images, duplicating them into the same destination file.
   4. Align the images to reconstruct the barrel field, ensuring proper anatomical order. Use the Move tool to position each layer manually, then click Edit |Transform |Rotate or Flip tool to adjust the orientation if needed. To fine-tune the alignment, use Edit | Transform | Distort and move the corners of the image to match the anatomical structures independently. To make alignment easier, set the blending mode of the top layer to Overlay in the Layers Panel.
   5. Compare anatomical landmarks such as barrel outlines and blood vessels to ensure correct positioning. To save the stack, click on File |Save as, name the file, and choose its format (PSD or TIFF); then, click OK.
3. Review the multiple sections containing the barrel field sequentially, identifying visible barrels in each section.
4. Manually outline each visible barrel on an additional common layer, ensuring the reconstruction of the entire barrel field. To create an extra layer, click Layer |New |Layer, name the layer, and click OK. Select the Brush Tool from the toolbar and on the new layer, manually draw each visible barrel.
   1. For greater precision, use zoom and adjust the brush size as needed. To zoom in and out of the image, use the Zoom Tool or press Ctrl + "+" / Ctrl + "-". To change the brush size, right-click on the canvas while the Brush Tool is active and move the Size slider.
5. Identify the selected barrel based on this reconstruction.
6. Find the selected barrel under the light microscope based on the pattern of the barrel field (five rows, each with four to seven barrels) and the arrangement of blood vessels that indicate its location within the barrel field.
7. In each section with a visible barrel cortex, identify the selected barrel under a stereomicroscope and cut it out together with the adjacent one using a razor blade.
8. Remove the small piece from the section, glue it on a resin block, and put it away to dry. Later, trim the excess resin surrounding the tissue.
   NOTE: Make sure to place the piece of tissue flat on the surface of the resin.
9. Use an ultramicrotome and a diamond knife to cut ultrathin consecutive sections (3-6 sections, 65 nm thick) (Figure 3A), and collect on formvar-coated, single-slot nickel grids (2 x 0.75 mm).
10. Once dry, contrast samples with 2% uranyl acetate for 3 min and 0.03% lead citrate for 1.5 min.
    NOTE: To contrast, it is important to wash the grids a few times and leave them to dry completely after each step. Use syringes with filters for uranyl acetate and lead citrate.
11. Take consecutive images from the hollow of the selected barrel (for example, B2) under a TEM (80 kV accelerating voltage at 4K or 8K magnifications, depending on the area and volume required for analysis.
    NOTE: Usually, at least a total volume of 100 µm³ per animal is analyzed.

4. Image analysis

NOTE: Creating and aligning stacks of TEM images can be done in the same way using both free software (like GIMP) and commercial one (e.g., Photoshop). Reconstructions, too, can be performed in the same way using both commercial software (e.g., 3D Studio Max) and completely free programs, such as Blender.

1. To estimate the density of synapses, stack a minimum of three TEM images in the graphics software (Figure 6). The greater the number of serial images, the higher the accuracy of the calculations.
   1. To open images, click File |Open, and select the images using Shift |Open; the images will open in separate windows.
   2. For each image, right-click on Layer 1 |Duplicate Layer. In the dialog box, enter the image number as the name, choose one image as the destination, and click OK. Repeat the same steps for all the images and duplicate them in the same destination.
   3. To save the stack, click File |Save as, name the file, choose its format (PSD or TIFF), and click OK.
2. To align TEM images, select the top image layer, and set the blending mode to Overlay in the Layers Panel. Use Edit |Transform |Rotate or the Flip tool to adjust the image orientation. Use the Move Tool to align structures.
   1. For more precise adjustments, use Edit |Transform | Distort to move the corners of the image independently. To increase precision, hold Ctrl and drag individual corner points. Use changes in the size and shape of mitochondria and axonal myelin sheaths between successive TEM images to verify the correct sequence of the images.
3. Use a dissector method for counting synapses.
   1. Add layer to the EM image stack by clicking Layer |New |Layer |OK. Ensure the new layer is transparent (with no background fill).
   2. To define the analysis area, click Layer |New |Layer |OK; then select the Rectangle Tool, go to Layer |Layer Style |Stroke, set stroke color, thickness (Size), and position (Inside recommended), click OK, and draw a rectangle. Select Fill: 0%. Mark all synapses present in the TEM images that do not cross two adjacent selected edges of the rectangle (e.g., the right and bottom or the top and left).
   3. Add a new layer for annotations by clicking Layer |New |Layer |OK. Use the Brush Tool to mark synapses on the new layer. Use the Zoom Tool and Hand Tool for precise marking.
   4. Choose one color to mark excitatory synapses and a different one for inhibitory synapses. To change the color, click on the foreground color box in the toolbar, select a color in the Color Picker window, and click OK.
4. If TEM images contain large structures (e.g., myelinated nerve fibers, large dendrites, or parts of the cell body), use a grid with multiple squares. This allows for the exclusion of any squares that are fully occupied by large structures.
   1. Add a new transparent layer by clicking Layer |New |Layer |OK. Enable grid view by clicking on View |Show |Grid, and enable snapping to grid by clicking on View |Snap to |Grid.
   2. Set grid spacing by clicking on Edit |Preferences |Guides, Grid & Slices. Select the Line Tool, set mode to Shape or Pixels, set line weight (e.g., 1 px), and draw vertical and horizontal lines along the visible grid to cover the entire image area.
5. Define the "top" and "bottom" sections for each image stack. Count only synapses that are absent in the last image but present in the previous images. This allows for accurate identification of synapses without overestimating their number.
6. Use the scale on the image and calculate the volume of the sample (stack of TEM images). Measure the area where the synapses are counted and multiply by the number of TEM images in the stack and the thickness of the ultrathin sections.
7. Divide the number of synapses by volume to determine the synapse density per unit volume.
8. Use all TEM images containing a dendritic spine (usually 10-12 consecutive electron micrographs) to reconstruct a dendritic spine and the synapses it forms.
   1. Open the image stack in the graphics software. Select the Crop Tool, mark the area containing the dendritic spine in the stack, and press Enter. The selected area will be cropped in all visible layers. Make only one visible at a time, hide other layers using the eye icon in the Layers Panel.
   2. Save each cropped image separately by clicking on File |Save as, name the file, choose a format (JPG recommended), and click OK. Repeat for each layer in the stack.
      NOTE: Before transferring the images to a 3D reconstruction software, it is best to first mark the elements that are being reconstructed with colors.
   3. Add a new layer for annotation by clicking on Layer | New |Layer | OK.
   4. Set the layer opacity to 30-50% in the Layers Panel.
   5. Select the Brush Tool, choose a different color for each structure (e.g., dendritic spine, synapse), and mark the elements. Work only on the new layer to preserve the original image. When finished, merge the annotation layer with the image, right-click on the top layer, and merge Down. Repeat these steps for each TEM image containing a cross-section of the dendritic spine.
9. Transfer the TEM images with visible cross-sections of the dendritic spine to 3D reconstruction software. To delete the default Cube object, right-click on the object, and delete. To align the viewport perspective before uploading images, click Z on the Navigation Gizmo or press Numpad 7. To upload images, click Add |Image |Reference, select the first image, and double-click on it to confirm. Repeat for each image, ensuring that the viewport perspective remains consistent throughout.1
   NOTE: To rotate the view, hold the middle mouse button, and move the mouse or use the Navigation Gizmo. To pan the view, hold Shift + middle mouse button, move the mouse, or use Move the view.
10. Arrange the TEM images at the appropriate distance, taking into account the thickness of the ultrathin sections (Figure 4A). To position the last TEM image along the Z axis, click on the image, click Object Properties (right sidebar), and enter the value in Location Z. The images may now be distributed between the first and the last image. To distribute them, select all images, press N to open the N-panel, and click on the Edit tab (in the N-panel) | Dist Z.
    NOTE: Before making any adjustments, align the viewport perspective, and click Z on the Navigation Gizmo or press Numpad 7.
    The Distribute tool must be enabled to distribute the images. To activate it, go to Edit | Preferences | Get Extensions, search "Distribute", and click Install. Once enabled, the Distribute tool remains active across sessions and does not need to be reactivated, as long as the same version of the program is used.
11. Manually outline the shape of the dendritic spine on each image using a single curve per TEM image (Figure 4B,C).
    1. Make sure only the first TEM image is visible. In the Scene Collection, hide all other objects using the eye icon. To create a base curve for the outline, click Add |Curve |Circle. Move the BézierCircle object  hold G, move the mouse, and click to accept the changes.
    2. To match the size of the dendritic spine profile, scale the object, hold S, move the mouse, and click. Enter the Modeling tab with the BézierCircle object selected. Align the viewport perspective again and click Z on the Navigation Gizmo or press Numpad 7.
    3. To adjust the shape of the curve, select a control point, hold G, move the mouse, and click.
    4. To adjust the direction and strength of the handles, select a handle point, press G to move or S to scale, and click to confirm. Repeat for each point.
    5. To add a new control point, select two existing control points, hold Shift and click each point, right-click, select Subdivide. When the curve corresponds to the boundary of the dendritic spine profile sufficiently, enter the Layout tab, right-click on the BézierCircle object and Duplicate Object, and click without moving the mouse. Create the same number of BézierCircle copies as there are TEM images used for reconstruction. Align them with the arranged TEM images using the Distribute tool as in step 4.10.
    6. Adjust each BézierCircle curve in the Modeling tab to match the dendritic spine profile in each image. Make sure only the currently edited BézierCircle object and corresponding TEM image are visible at a time.
       NOTE: All right-click actions required for certain steps must be performed within the scene boundary. The successive curves must correspond to the boundaries of consecutive dendritic spine sections and contain the same number of control points.
12. Connect adjacent lines using the 3D reconstruction function to form a continuous surface and create the final 3D shape (Figure 4D).
    1. Make sure only the BézierCircle objects are visible. In the Scene Collection, hide all other objects using the eye icon in the Sidebar.
    2. To create the dendritic spine object, go to Layout tab, select all BézierCircle objects, right-click, and click Join. Convert the joined object to a mesh by right-clicking and selecting Convert To |Mesh.
    3. To connect the edges of the joined object, enter the Modeling tab, press 2 to enter Edge Select mode |Select |All |right-click |Bridge Edge Loops.
    4. To fill the gap at the top the object, deselect all by clicking next to the object, hold Alt and click one edge section of the top ledge, right-click, select Extrude Edges, drag upward on the Z axis, press S and move the mouse to shrink the new edge loop, click to accept the change. While the new edge loop is still selected, right-click and select New Face from Edges. Fill the bottom gap using the same method.
    5. To smooth the object, enter the Layout tab, right-click, and select Shade Smooth. To further smooth the shape, go to Modifiers (right sidebar) |Add Modifier |Generate |Subdivision Surface. Set both the Viewport and Render levels to the same value (recommended 1).
13. Repeat steps 4.11-4.12 for synapse reconstruction (Figure 4E,F).
14. Assign a color to each structure and use the rendering function to prepare the final image (Figure 4G,H). Go to Viewport Shading menu and select Rendered to preview material changes. Select the object to be colored and click Material (right sidebar) |New |adjust Base Color.
    1. To refine the material appearance, adjust Roughness or Metallic sliders as needed. Use this method to assign a distinct color to each structure.
    2. To adjust scene lighting, click on the Light object, hold G, and move the mouse to reposition it. To change light intensity, go to Data (right sidebar) and adjust the Power value.
    3. To position the Camera, orient the viewport to face the reconstructed spine as it should appear in the final image, press Ctrl + Alt + Numpad 0. Select the visible Camera frame, hold G, and move the mouse to reposition it if necessary.
    4. To prepare for rendering, go to the Output (right sidebar), change Resolution X and Y values to match the desired image. To improve image quality, go to Render (right sidebar) |Render, adjust Samples value to between 120 and 250. For a transparent background, go to Render |Film |check Transparent. To render the image, press F12 or go to Render (upper menu) |Render Image.

Results

To properly apply our method to the analysis of circadian synaptic changes, it is necessary to begin by selecting at least four time points at equal intervals, with two points for each phase of the animals' activity (every 6 h). At these designated time points, the animals are sacrificed, and their brains are collected. This approach allows for the identification of daily or circadian patterns of synaptic plasticity and links them to changes in the animals' locomotor activity (Figure 1). Usi...

Discussion

Here, we presented the methodology used for studying the circadian plasticity of synapses and the reconstruction of dendritic spines in the barrel cortex of mice. To ensure reliable results, the circadian plasticity study should include at least four time points. Our research showed that data from two time points -- one during the rest phase (day) and one during the activity phase (night) -- provided information on the daily differences between the activity phases of animals under LD 12:12 conditions. Total synapse densi...

Materials

List of materials used in this article

Name	Company	Catalog Number	Comments
Aclar film	Agar scientific	AGL4458
Blender software			version: v. 2.91.2
Cacodylic acid, sodium salt trihydrate	Polysciences	01131-100
Concave slide	Menzel	9.161 151
Diamond knife	Diatome	15-US	Knife angle: Ultra 45°
Digital camera	Nikon	DXM 1200 F
di-Sodium Hydrogen Phosphate Dodecahydrate	POCH	799280115
Embedding resin	Polysciences	08792-1	Luft formulation
Ethanol	Pol-Aura	PA-11-0004
GIMP software			version: v 3.0.4
Lead citrate	TAAB	L018
Light microscopy	Nikon	Optiphot
Osmium tetroxide	Polysciences	0223C-10	Crystalline (99.95%)
Paintbrush
Photoshop software			version: CS5
Potassium ferricyanide	Sigmaaldrich	P8131
Propylene oxide	Sigmaaldrich	82320	puriss. p.a., ≥99.5% (GC)
Single slot grids	Agar scientific	AGG2525	2 x 0.75 mm or 2 x 1 mm
Sodium chloride	Chempur	117941206
Sodium dihydrogen phosphate dihydrate	Chempur	117991808
Stereo microscope	pzo
Syring filter	Biosens	BS25PES045
TEM	Joel	JEM-2100
Ultramicrotome	Leica	UC7
Uranyl acetate	Lachema
Vibratome	Leica	VT1000 S